Estimation question

How many reviews are left on Amazon on a given Monday?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Funnel-based estimation: converting total order volume into a review count using a realistic review-leave rate.

Estimate Amazon's daily US order volume from a rough annual figure divided across the year, adjusted slightly for a weekday like Monday.
Estimate what fraction of orders result in a delivered item eligible for review by Monday, accounting for shipping lag, so use orders delivered, not placed, that day.
Apply a realistic review-leave rate, commonly cited around 1-5% of purchases, higher for some categories.
Multiply delivered-eligible orders by the review rate to get total reviews left that day.
Sanity check the order of magnitude against the known scale of reviews visible across Amazon's catalog.

Distinguishes orders placed from orders delivered, since reviews follow delivery with a lag.
States a specific, labeled review-leave-rate assumption instead of an arbitrary final number.
Arrives at a plausible order of magnitude, low hundreds of thousands per day, and sanity-checks it.

Using purchase volume directly without accounting for shipping lag before a review can be left.
Not stating the review-leave-rate assumption, the single biggest driver of the final number.

How would incentivized reviews or Vine reviewers change this estimate?
How would the review rate differ between categories like electronics and groceries?